Past studies have investigated the relationship of topography to above-ground forest carbon density and water availability in the Barro Colorado Island 50 ha plot. Terrain slope was found to be the strongest predictor for carbon density across BCI compared with geologic substrate and forest age (Mascaro et al. 2011). Species richness of saplings at small spatial scales was higher on slopes (Hubbell et al. 1999). Water availability was found to be higher on slopes than plateaus during the dry season (Becker et al. 1988) which could hypothetically create a more suitable habitat for seedlings and consequently increase species richness (Hubbell et al. 1999). Variation in how heterospecifics respond to topography related resource availability (Harms et al. 2001; Engelbrecht et al. 2007; Comita et al. 2009) may cause spatial variation in alpha diversity. We quantified spatial variation in alpha diversity at multiple spatial scales ranging from 25 m2 to 62,500 m2. We quantified topographic properties of the Barro Colorado Island 50 ha plot forest with discrete-return light detection and ranging (DRL) data to investigate how variation in topography relates to spatial heterogeneity of alpha diversity. Using high-resolution DRL we calculated elevation, terrain slope, rugosity and corrected three-dimensional sub-plot area on the ground at each spatial scale. There have been no published attempts to use DRL-derived topography to study species richness and tree diversity in the 50-ha plot. This is the first study to our knowledge to investigate the scale dependence of species richness and alpha diversity on topography and forest structure in a tropical forest.
Results/Conclusions
Patterns of both topography and alpha diversity change depending on spatial scale. Species richness and alpha diversity appear to be strongly correlated with terrain slope and rugosity at coarse spatial scales and are correlated with stem density at fine spatial scales. Species richness and alpha diversity are strongly coupled with the number of individuals at fine spatial scales, but become uncoupled at coarse spatial scales where topographic variables appear to be more important. The predictive power of stem density on species richness is illustrated by the decreasing R-squared values for subplot sizes 5, 10, 20, 50, 100, which decrease from .71, .50, .35, .20 and .07, respectively. At coarser spatial scales (+50 m) topographic variables such as elevation, slope, corrected area and rugosity are the best predictors of both diversity and evenness.